Average (Mean) Calculator

Which Average Should You Use?

'Average' isn't one thing — it's a family of measures, and choosing the right one matters because they can give very different answers from the same data. The mean (arithmetic average) sums all values and divides by the count; it's best for symmetric, normally distributed data without extreme outliers, and is what most people mean by 'average'. The median is the middle value when data is sorted (or the average of the middle two for an even count); it's far more robust to outliers because extreme values don't shift it much, which is why it's used for income, house prices, and other skewed distributions. The mode is the most frequently occurring value; it's useful for categorical data ('most common shoe size sold') and for data with a clear peak. The geometric mean (the nth root of the product of n values) is used for growth rates, returns, and anything multiplicative — averaging returns of +50% and −50% via the arithmetic mean wrongly gives 0%, while the geometric mean correctly captures the actual outcome (you'd be down 25%). The harmonic mean handles rates and ratios (the correct average of 60 mph for one leg and 30 mph for the next is the harmonic mean, 40 mph, not the arithmetic 45). Picking the wrong measure misleads, sometimes badly — which is why news stories quoting 'average' should usually be checked for which average they mean.

When Mean Misleads?

The most common 'average' problem is when one or two extreme values pull the mean far from where most data actually sits. Worked example: if 9 people earn £20,000 and 1 person earns £200,000, the mean is (9×20,000 + 200,000) ÷ 10 = £38,000 — well above what 90% of the group actually earns. The median income (£20,000) tells the real story of typical experience. This is exactly why median household income is the standard measure for cost-of-living and inequality discussions, not the mean: a few very-high earners would otherwise inflate the figure to something most households would never see. Skewed distributions are everywhere: house prices (a few mansions skew the mean upward), salaries within companies, wealth distribution, response times on websites, and pretty much any dataset with a long tail. A useful diagnostic: if the mean is much higher than the median, your data is right-skewed (extreme high values); if much lower, left-skewed. When the mean and median agree, the data is roughly symmetric. Outliers in either direction warrant attention — sometimes they're errors to remove, sometimes they're real and important. The mean is mathematically convenient (it has nice properties for statistics) but the median is often more honest for describing what a 'typical' value is. A good practice for important data: report both, plus the spread, so readers can see the full picture rather than be misled by a single number.

What's the key thing to understand about Standard Deviation?

An average tells you the centre of your data; standard deviation tells you the spread around it — and you usually need both for the picture to make sense. Two datasets can have the same mean but very different distributions: scores of {49, 50, 51} and {0, 50, 100} both average 50, but one is tight and the other wildly spread. Standard deviation captures this: roughly the average distance of values from the mean. A low SD means data is clustered closely around the mean (consistent); a high SD means it's spread widely (variable). For roughly normal distributions, useful rules: about 68% of values fall within one SD of the mean, 95% within two SDs, and 99.7% within three SDs (the '68-95-99.7 rule' or empirical rule). This is why exam grading can use SD-based curves, and why quality control flags values more than 3 SDs from target as likely defects. Practical applications: in finance, SD of returns measures investment volatility (higher SD = higher risk); in manufacturing, six-sigma quality means defects are over 6 SDs from the mean (extraordinarily rare); in science, error bars on graphs usually represent one SD. There are two slightly different SD calculations: population SD (dividing by n) for the whole population, and sample SD (dividing by n-1) when your data is a sample estimating a wider population — for most everyday calculations on a sample, use n-1. Together, mean and standard deviation provide a much fuller summary than either alone.

What's the key thing to understand about Weighted Averages?

A simple average treats every value equally, but in real life some values matter more than others — and that's when a weighted average is the right tool. Each value is multiplied by its weight (how much it counts), the weighted values are summed, and the total is divided by the sum of the weights. Worked example: a student scores 65% on coursework worth 30% of the grade and 80% on an exam worth 70%. The simple average is (65 + 80) / 2 = 72.5%, but the correct weighted average is (65 × 30 + 80 × 70) / 100 = (1,950 + 5,600) / 100 = 75.5%. The simple average understates the final result because it ignores that the exam carries more weight. Weighted averages appear constantly: degree classifications combine module marks weighted by credits; share-price indices like the FTSE 100 weight companies by market value (so movements in larger companies count more); a grade-point average weighs each grade by the credit hours of that course; statistical surveys weight responses to correct for over- or under-represented groups; and portfolio returns combine each asset's return weighted by its share of the portfolio. The general rule: when the things being averaged have different importance, sizes, or frequencies, the weighted average is what you want. Simple averages of percentages from groups of different sizes are particularly prone to misleading results — always check whether weights should be applied. This calculator can handle weighted averages alongside the simple cases.

When Is the Arithmetic Mean the Wrong Average to Use?

Two lesser-known averages solve specific problems the arithmetic mean gets wrong: the geometric mean for growth rates and ratios, and the harmonic mean for rates like speed where the same distance is covered at different speeds.

Worked example: a car drives 60 miles at 30mph, then the same 60 miles at 60mph. The arithmetic mean of the two speeds is (30+60)/2 = 45mph — but the car actually spent 2 hours on the first leg and 1 hour on the second, covering 120 miles in 3 hours, for a true average speed of 40mph. The harmonic mean of 30 and 60 — 2÷(1/30 + 1/60) = 40 — matches the real answer exactly, while the arithmetic mean overstates it, because it doesn't account for the slower leg taking proportionally more time.

Worked Example: Comparing Mean and Weighted Mean for a Grade

A student's coursework scores are 70, 85, and 90 on three assignments worth equal credit. The simple arithmetic mean is (70 + 85 + 90) / 3 = 245 / 3 ≈ 81.67. Now suppose the course weights these assignments differently: the first is worth 20% of the grade, the second 30%, and the third 50%. The weighted mean is (70 × 0.20) + (85 × 0.30) + (90 × 0.50) = 14 + 25.5 + 45 = 84.5. The weighted result is higher because the student's strongest score (90) carries the most weight, while the simple mean treats all three assignments as equally important and understates the final grade. This is exactly why course syllabi specify weighting schemes rather than relying on a plain average -- the weighted mean reflects how much each component actually counts toward the outcome.

Mean, Median, and Mode Side by Side

Each measure of central tendency answers a slightly different question, and choosing the wrong one for a dataset can misrepresent it.

A well-known illustration: if nine people earn $40,000 and one earns $2,000,000, the mean income is $236,000 -- a figure that misrepresents what a typical person in the group earns -- while the median remains $40,000, accurately reflecting the center of the group. National income and house-price statistics are reported as medians for exactly this reason.